SQL Server Performance

Architecture for High Availability and Disaster Recovery

Discussion in 'SQL Server 2008 General DBA Questions' started by ismailadar, Jan 6, 2011.

  1. ismailadar New Member

    Hi all,
    i need your valuable offers for architecture that i plan to install. I have 3 servers and i need to developped an architecture for high availabilty and disaster recovery scenarios. It should desing with minimum cost. Waht is your scenarios for high availabilty and disaster recovery with 3 servers?
    Thanks in advanced
  2. adityaguthey New Member

    We use peer to peer replication between three servers located in three dataCenters. This is a high availability and DR solution
  3. satya Moderator

    In terms of cost you are referring what is the limit on budget?
    As you are using SQL Server 2008 there are plenty of options available based on the Editions that you are using. If its STANDARD edition then you are limited to Database Mirroring & Log shipping (to some extent of replication too).
    As suggested above Peer to peer is an Enterprise edition feature, if you are looking for HA & DR then Clustering is your best bet here too.
    Before I go into the best feature for HA & DR, there are certain levels of questions that you need to ask yourself before planning to design a concept:
    What is the key criteria for providing HA : Maximum availability?
    What is the key criteria for providing DR : near-immediate recovery?
    Do you consider performance for your application to be incorporated as a HA process?
    Do you eliminate or raise a risk for any single-point failure on the architecture?
    Are there any thought-process procedures and practices?
    What is the minimum uptime required for your application?
  4. ismailadar New Member

    thanks for reply we use SQL server 2008 enterprise editon so we can able to get more solution beyond Database Mirroring & Log shipping.We sholul also provide HA and DR simultaneously. Also i thinmk clustering is best for us but unfortunatley we do not have san. So how can we implement a HA and DR scenario except using clustering? Another question is how can we use Database Mirroring & Log shipping together.?
  5. satya Moderator

    As you are using EE but not the SAN its right that you will not able to utilize the cluster in this case.
    Database Mirroring can be used for the High Availability provision where the databases can be mirrored between the sites, with AUTOMATIC failover or manual failover as per your application SLA. In addition to DB mirroring log shippign can also be implemented using the normal procedures where the third server can be used to as SECONDARY server for LS purpose.
  6. RamJaddu Member

Share This Page